To the Editor.— We are delighted that Drs Shuman and Selednik have sought to identify the risk factors of periventricular leukomalacia (Archives1980;37:231-235). Some of their findings are similar to ours, but others are not.1 The subjects of our study were the 560 newborns in the Collaborative Perinatal Project who were born alive, died within the first 28 days, and had postmortem examination of the brain. Of the more than 1,200 variables we explored, 96 had a distribution among the 57 cases that differed significantly (P<.1) from the distribution among the 503 controls. These were entered into a Seigel-Greenhouse linear discrimination model that allowed estimation of the relative risk for each variable when all the other selected variables were considered. We found that the risk of periventricular leukomalacia was prominently increased in those born before the 36th week of gestation (relative risk estimate [RR] = 4.2; 95% confidence range